Loading…
EyeleveN Product Documentation
Introduction, architecture, operating model, and launch checklist to run EyeleveN as a production LATAM AI platform.
1) What EyeleveN Is
EyeleveN is a multi-tenant AI Digital Employee operating system, not a chatbot widget. Businesses deploy AI employees that run lead qualification, booking, follow-up, support triage, and CRM updates through WhatsApp-first workflows.
2) Core Architecture
Channel Gateway -> Webhook Ingestion -> Conversation Orchestrator -> LLM + Tool Calling -> Workflows -> Knowledge/RAG -> Human Takeover Console -> Analytics/Billing.
3) Multi-Tenant Foundation
Each workspace has isolated channels, AI employees, conversations, documents, workflows, integrations, analytics, and plan limits from day one.
4) Human + AI Operations
Teams can switch AI/Human mode, pause automation, assign owners, save internal notes, and escalate with audit logs.
5) Safety and Control
Idempotent webhooks, signature verification, role-based access control, policy guardrails, and action audit trails reduce operational risk.
Go-Live Checklist
Connect real WhatsApp Cloud API credentials and verify webhook endpoint.
Apply all Supabase migrations and sync the real production tenant + integrations.
Configure billing plans, checkout flow, and subscription lifecycle webhooks.
Set up queue worker cron (job draining) and error alerting.
Define SLAs for human takeover windows and escalation routing.
Upload real SOPs, pricing, policies, and catalogs in Knowledge Base.
Enable observability: logs, uptime, queue depth, failed workflow alerts.
Prepare LATAM GTM assets: onboarding, case studies, WhatsApp templates.
Monitoring Endpoints
/api/health
Public health signal. Send x-eyeleven-health-secret (or Bearer token) to receive detailed diagnostics.
/api/ready
Public readiness probe. Expected: HTTP 200 when ready, HTTP 503 when not ready. Includes critical/optional check summary; detailed checks still require health secret header.
Legal and Compliance